A method and an apparatus for bi-layer segmentation of an image or a sequence of images are described. A classifier is derived based on depth data of the image and another classifier is derived based on color data of the image. The image is then segmented by maximizing a weighted sum of matching scores derived from the classifiers based on depth data and color data of the image. The classifier based on color data of the image is derived using color sampling subsequent to generating an initial segmentation of the image.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method for bi-layer segmentation of an image, the method comprising deriving a classifier based on depth data of the image; deriving a classifier based on color data of the image; and segmenting the image by maximizing a weighted sum of matching scores derived from the classifiers based on depth data and color data of the image.
2. The method according to claim 1 , wherein the classifier based on color data of the image is derived using color sampling.
3. The method according to claim 1 , further comprising generating an initial segmentation of the image.
4. The method according to claim 3 , wherein the initial segmentation of the image is generated based on depth data associated to the image.
5. The method according to claim 4 , wherein the initial segmentation of the image is generated by performing a statistical analysis of a depth distribution of the image or further images of an image sequence comprising the image.
6. The method according to claim 1 , wherein the classifier based on depth data is constructed using a statistical distribution of the depth data across the image and computing a cumulative likelihood of depth measurements.
7. The method according to claim 1 , wherein the classifier based on color data is derived using color sampling, where matching scores are given by a distance in color space between each pixel and a closest sample within a set of samples randomly selected from a specific area of the image.
8. The method according to claim 1 , wherein maximizing the weighted sum of matching scores is performed using a discrete optimization technique.
9. The method according to claim 8 , wherein the discrete optimization technique is graph cut.
10. A method for bi-layer segmentation of an image, the method comprising generating an initial segmentation of the image;—deriving a classifier based on color data of the image using color sampling; and segmenting the image by maximizing a sum of matching scores derived from the classifier based on color data of the image.
11. The method according to claim 10 , wherein the initial segmentation of the image is generated based on depth data associated to the image.
12. The method according to claim 11 , wherein the initial segmentation of the image is generated by performing a statistical analysis of a depth distribution of the image or further images of an image sequence comprising the image.
13. The method according to claim 10 , wherein the classifier based on color data is derived using color sampling, where segmentation scores are given by a distance in color space between each pixel and a closest sample within a set of samples randomly selected from a specific area of the image.
14. The method according to claim 10 , wherein maximizing the sum of matching scores is performed using a discrete optimization technique.
15. The method according to claim 14 , wherein the discrete optimization technique is graph cut.
16. An apparatus for bi-layer segmentation of an image, wherein the apparatus is configured to perform: deriving a classifier based on depth data of the image; deriving a classifier based on color data of the image; and segmenting the image by maximizing a weighted sum of matching scores derived from the classifiers based on depth data and color data of the image.
17. The apparatus according to claim 16 , wherein the apparatus is configured to derive the classifier based on color data of the image using color sampling.
18. The apparatus according to claim 16 , wherein the apparatus is further configured to generate an initial segmentation of the image.
19. The apparatus according to claim 18 , wherein the apparatus is configured to generate the initial segmentation of the image based on depth data associated to the image.
20. The apparatus according to claim 19 , wherein the apparatus is configured to generate the initial segmentation of the image by performing a statistical analysis of a depth distribution of the image or further images of an image sequence comprising the image.
21. The apparatus according to claim 16 , wherein the apparatus is configured to construct the classifier based on depth data using a statistical distribution of the depth data across the image and computing a cumulative likelihood of depth measurements.
22. The apparatus according to claim 16 , wherein the apparatus is configured to derive the classifier based on color data using color sampling, where matching scores are given by a distance in color space between each pixel and a closest sample within a set of samples randomly selected from a specific area of the image.
23. The apparatus according to claim 16 , wherein the apparatus is configured to maximize the weighted sum of matching scores using a discrete optimization technique.
24. The apparatus according to claim 23 , wherein the discrete optimization technique is graph cut.
25. An apparatus for bi-layer segmentation of an image, wherein the apparatus is configured to perform: generating an initial segmentation of the image; deriving a classifier based on color data of the image using color sampling; and segmenting the image by maximizing a sum of matching scores derived from the classifier based on color data of the image.
26. The apparatus according to claim 25 , wherein the apparatus is configured to generate the initial segmentation of the image based on depth data associated to the image.
27. The apparatus according to claim 26 , wherein the apparatus is configured to generate the initial segmentation of the image by performing a statistical analysis of a depth distribution of the image or further images of an image sequence comprising the image.
28. The apparatus according to claim 25 , wherein the apparatus is configured to derive the classifier based on color data using color sampling, where segmentation scores are given by a distance in color space between each pixel and a closest sample within a set of samples randomly selected from a specific area of the image.
29. The apparatus according to claim 25 , wherein the apparatus is configured to maximize the sum of matching scores using a discrete optimization technique.
30. The apparatus according to claim 29 , wherein the discrete optimization technique is graph cut.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
March 4, 2013
August 11, 2015
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.